Adipocytes isolated from apolipoprotein E (apoE)-knockout (EKO) mice display alterations in triglyceride (TG) metabolism and gene expression. The present studies were undertaken to evaluate the impact of endogenously produced adipocyte apoE on these adipocyte parameters in vivo, independent of the profoundly disturbed metabolic milieu of EKO mice. Adipose tissue from wild-type (WT) or EKO mice was transplanted into WT recipients, which were then fed chow or high-fat diet for 8–10 wk. After a chow diet, freshly isolated transplanted EKO adipocytes were significantly (P < 0.05) smaller (70%) than transplanted WT adipocytes and displayed significantly lower rates of TG synthesis and higher rates of TG hydrolysis. Transplanted EKO adipocytes also had higher mRNA levels for adiponectin, perilipin, and genes coding for enzymes in the fatty acid oxidation pathway and lower levels of caveolin. After a high-fat diet and consequent increase in circulating lipid and apoE levels, transplanted WT adipocyte size increased by 106 × 103 μm3, whereas EKO adipocyte size increased only by 19 × 103 μm3. Endogenous host adipose tissue harvested from WT recipients of transplanted WT or EKO adipose tissue did not demonstrate any difference in adipocyte size. Consistent with the in vivo observations, EKO adipocytes synthesized less TG when incubated with apoE-containing TG-rich lipoproteins than WT adipocytes. Our results establish a novel in vivo role for endogenously produced apoE, distinct from circulating apoE, in modulation of adipocyte TG metabolism and gene expression. They support a model in which endogenously produced adipocyte apoE facilitates adipocyte lipid acquisition from circulating TG-rich lipoproteins.
- adipose tissue
the increased metabolic and cardiovascular risk that accompanies the expansion of adipose tissue is a major health problem as the prevalence of obesity rises in the United States (12, 15, 22). Adipose is a metabolically active tissue involved in internalizing and storing energy delivered by circulating triglyceride (TG)-rich lipoproteins (TGRL). The fatty acids derived from the hydrolysis of TG carried by these lipoproteins are reesterified to form intracellular TG and stored within intracellular lipid droplets, where they can subsequently be released into the circulation in response to metabolic and hormonal signals (6, 9). These fatty acids can then serve as substrates for oxidation or, alternatively, as substrates for the synthesis of TG in distant tissues such as liver and muscle. In addition to modulating the systemic metabolism of TG and fatty acids, adipocytes and adipose tissue secrete numerous humoral factors that modulate gene expression and metabolic pathways in distant tissues (20). In obesity, systemic TG turnover, fatty acid flux, and the expression and secretion of these adipocyte-derived factors are altered (2, 8).
The transplantation of adipose tissue has proved to be a useful model for providing physiological context for the role of adipose tissue in systemic metabolism. For example, surgical implantation of adipose tissue has been shown to reverse diabetes in lipoatrophic mice (11). The heterogeneity of fat from different adipose tissue depots has been established by recent studies demonstrating divergent metabolic effects of the transplantation of visceral or subcutaneous fat (30). A similar experimental transplantation paradigm has demonstrated the divergent effects of visceral and subcutaneous fat on large-vessel atherosclerosis in atherosclerosis-prone mice (28). Adipose tissue transplantation between inbred strains of genetically engineered mice has also been a useful tool for providing insight into the significance of specific adipocyte genes for systemic metabolism and adipocyte function (4).
Adipocytes produce a number of proteins at a high level that are also produced by other tissues and are present at a significant concentration in blood (20). One of these is apolipoprotein E (apoE) (34). High-level expression of apoE is evident in adipocytes and other cells that experience high lipid flux, such as macrophages (21, 26) and hepatocytes (24). In the latter cell type, apoE is produced and secreted as a surface component of TGRL particles, and hepatocytes can contribute >90% of circulating apoE (24). An important role for apoE carried by circulating TGRL in adipocyte metabolism has been established by investigations from multiple laboratories (5, 14, 19). For example, adipocytes from apoE-knockout (EKO) mice are smaller than those from wild-type (WT) mice (19). Systemic apoE deficiency also impairs TG clearance and produces resistance to obesity in strains of genetically obese mice (14). It has also been shown that apoE on VLDL is required for VLDL-induced lipogenesis in cultured adipocytes (5) and that adenoviral expression of apoE in genetically obese apoE-deficient mice increases the size of brown and white adipose tissue adipocytes (10).
On the other hand, as noted above, apoE is highly expressed in adipocytes, and in vivo observations using EKO mice do not discriminate between the impact of apoE in circulating TGRL and that of endogenously produced adipocyte apoE for modulating adipocyte lipid metabolism. Using isolated and cultured adipocytes from EKO mice, we previously reported that endogenously expressed apoE can have a significant impact on adipocyte TG and fatty acid metabolism in vitro (19). For example, the adenoviral expression of apoE increases TG and fatty acid content of cultured EKO adipocytes, and during 24- to 48-h incubations with apoE-rich TGRL, less TG is accumulated in EKO than in WT adipose tissue. These in vitro observations suggest an important role for endogenously produced apoE, independent of apoE in circulating TGRL, for enhancing adipocyte TG content and metabolism. However, somewhat contradictory results have recently been reported by Carmel et al. (3), who showed that high expression of apoE impairs lipid storage in the adipocyte line SW872.
We used the adipose tissue transplantation model to evaluate a specific role of endogenously produced apoE for modulating adipocyte size, lipid metabolism, and gene expression within a WT in vivo context, independent of significant alterations in circulating lipid or apoE levels. Adipose tissue from EKO or WT donor mice was transplanted into WT recipients. After 8–10 wk on a chow or high-fat diet (HFD), adipose tissue was harvested for analysis.
MATERIALS AND METHODS
Cell culture medium was purchased from Invitrogen (Carlsbad, CA), all chemicals from Sigma (St. Louis, MO), organic solvents from Thermo Fisher Scientific (Pittsburgh, PA), [14C]oleate from PerkinElmer (Wellesley, MA), and total cholesterol (TC) and TG assay kits from Wako Chemicals USA (Richmond, VA). All other materials were obtained from previously identified sources (16, 17).
Adipose tissue transplantation.
All animal protocols and surgical procedures were approved by the Institutional Animal Care and Use Committees of the University of Illinois at Chicago. C57BL/6J control mice were purchased from Charles River (Wilmington, MA). EKO breeder pairs on a C57BL/6J background were purchased from Jackson Laboratories (Bar Harbor, ME) and bred in-house. Adipose tissue transplantation was performed as described previously with minor modifications (11). Briefly, epididymal white adipose tissue was isolated from 14-wk-old male WT or EKO donor mice and rinsed in cold PBS. All recipient mice were 12-wk-old male WT mice. Recipients were shaved, and four 5-mm skin incisions were made on the back. WT or EKO adipose tissue grafts (400 mg total per mouse) were inserted subcutaneously in equal portions into the four subcutaneous dorsal skin incisions, which were closed with surgical clips. After transplantation, mice were housed individually and fed chow (13.4% kcal from fat, diet no. 7912; Harlan Teklad, Madison, WI) or HFD (60% kcal from fat, diet no. TD6414; Harlan Teklad). Mice were euthanized 8–10 wk after transplantation for collection of nonfasting plasma and transplanted adipose tissue.
Adipocyte isolation and sizing and total lipid estimation.
Mature adipocytes were isolated from adipose tissue as previously described in detail (19). For sizing, images of suspended adipocytes were recorded using a light microscope, and adipocyte size was measured using an automated analysis program (CCAP, Mayo Clinic, Rochester, MN) (29). A total of 800–1,000 adipocytes in suspension were counted per recipient mouse, and there were five recipient mice in each experimental group per experiment. Total cellular lipids were estimated by CCAP software using a previously published method based on adipocyte volume and the density of triolein (29). For microscopy, transplanted fat was fixed, embedded in paraffin, and stained with hematoxylin-eosin, as previously described (19). Preadipocytes from the adipose tissue stromovascular fraction were isolated and differentiated into adipocytes by 3 days of incubation in insulin, dexamethasone, and isobutylmethylxanthine, as previously described in detail (19). These cells were used for experiments 10 days after completion of the incubation in the differentiation cocktail.
TG synthesis and hydrolysis.
For measurement of TG synthesis, freshly isolated mature adipocytes were incubated with [14C]oleate-BSA complex (0.25 μCi/ml, specific activity 70,000 dpm/mol) in DMEM for 2 h, as previously described in detail (19). Labeled cells were washed and extracted using Folch solution, and labeled TG was separated by thin-layer chromatography in a solvent system of hexane-ethyl ether-acetic acid (90:30:1), as previously described. TG spots were scraped, and radioactivity in spots was measured in a scintillation counter.
For measurement of adipocyte TG synthesis in response to incubation with apoE-containing TGRL, human VLDL was isolated by sequential density ultracentrifugation as previously described (19); apoE content of isolated VLDL has been previously demonstrated (19). VLDL (100 μg/ml) was incubated with freshly isolated mature adipocytes or cultured adipocytes in 0.1% BSA-DMEM containing [14C]glucose (0.5 μCi/ml) for 6 h at 37°C. After the incubation, cells were washed and lipids were extracted for measurement of labeled TG (see above).
TG hydrolysis rate was estimated by measurement of glycerol release into the medium from isolated mature adipocytes over 90 min. Isoproterenol-stimulated TG hydrolysis was measured after incubation of the cells with or without 50 μM isoproterenol over 120 min. Glycerol released in the medium was measured using a free glycerol determination kit (Sigma). TG synthesis and hydrolysis rate were normalized to adipocyte cell number by correction of values for adipocyte DNA content.
Quantitative PCR and Western blot.
Total RNA was extracted from adipocytes with use of an RNeasy mini kit (Qiagen, Valencia, CA). Quantitative real-time PCR analysis of gene expression was performed using a Stratagene MX 3000P with Brilliant SYBR Green QRT-PCR Master Mix (Stratagene, La Jolla, CA). Some primers used in this study have been published previously (19). Additional primers are as follows: adipocyte triglyceride lipase [GATGACCACCCTTTCCAACA (forward) and TGGCCCTCATCACCAGATAC (reverse)], hormone-sensitive lipase [GGGAGCACTACAAACGCAAC (forward) and CAGAGACGACAGCACCTCAA (reverse)], and lipoprotein lipase [TATCCCAATGGAGGCACTTTC (forward) and CTGTATGCTTTGCTGGGGTTT (reverse)]. Quantitative PCR data were evaluated using the comparative critical threshold (Ct) method and normalized to the average of the endogenous control gene β-actin. Changes in gene expression were calculated by 2. Plasma apoE levels were estimated by quantitative Western blot using purified apoE prepared from VLDL to generate a standard curve.
DNA and protein estimations.
DNA was isolated from adipocytes with a DNeasy kit (Qiagen) according to the manufacturer's instructions. DNA mass was measured using a Pico green DNA kit (Molecular Probes, Invitrogen). Cell protein was measured using a DC protein kit (Bio-Rad, Hercules, CA).
Values are means ± SD of five mice per experimental group. Statistical differences were analyzed using Student's t-test or ANOVA (SPSS 15.0, SPSS, Chicago, IL). P < 0.05 was considered significant.
It has previously been shown that EKO mice have lower body weight, lower total body fat mass, and smaller adipocytes than age- and gender-matched WT mice, despite their marked hyperlipidemia (5, 10, 19). Adipose tissue and adipocytes isolated from EKO mice are exposed to a markedly disturbed in vivo environment that, importantly, includes total absence of circulating apoE. It is impossible, therefore, to determine whether the presence of smaller adipocytes in EKO mice results from the absence of circulating apoE or the absence of endogenously produced adipocyte apoE. To address this question in a physiological in vivo context, we transplanted adipose tissue from EKO donors or from age- and gender-matched WT donors into WT recipients (designated EKO-WT and WT-WT recipient mice, respectively). The recipients were fed a chow diet for 8 wk, and transplanted adipose tissue was harvested and the adipocytes were isolated. At time of harvest, there was no significant difference in body weight between mice receiving WT or EKO adipose tissue, although recipients of EKO adipose tissue tended to weigh less (32.6 ± 2.0 vs. 30.5 ± 0.9 g). There was also no difference in food intake between the two groups of mice. Plasma free fatty acid levels were 0.47 ± 0.08 and 0.51 ± 0.09 mmol/l in recipients of WT and EKO adipose tissue, respectively (not significant). Glucose levels were not significantly different between the two transplantation groups (182 ± 19 and 175 ± 19 mg/dl in recipients of WT and EKO adipose tissue, respectively). A representative section of transplanted EKO or WT adipose tissue in Fig. 1A demonstrates the larger size of transplanted WT adipocytes. Figure 1B shows the cell volume frequency distribution of adipocytes isolated from EKO or WT transplanted adipose tissue harvested and pooled from five WT recipients in each group. Peak frequency volume was smaller in transplanted EKO adipocytes. Figure 1C shows adipocyte cell volume in transplanted EKO and WT adipose tissue harvested from 15 WT recipients in each transplant group. Adipocyte mean volume was significantly lower (by >70%) in adipocytes isolated from transplanted EKO adipose tissue after 8 wk in a normalized WT in vivo environment. The calculated adipocyte lipid content per cell was also lower (0.077 ± 0.021 vs. 0.022 ± 0.001 μg/cell, P < 0.05).
We previously showed that TG synthesis is lower and TG hydrolysis is higher in adipocytes that are freshly isolated from EKO mice than from WT controls (19). As noted above, the results of these previous experiments do not discriminate between the impact of a markedly abnormal in vivo environment and the absence of adipocyte apoE expression. We next measured TG synthesis and hydrolysis in adipocytes freshly isolated from transplanted adipose tissue. TG synthesis was significantly lower in transplanted EKO adipocytes (Fig. 2A), whereas TG hydrolysis was twofold higher (Fig. 2B). It has previously been shown that adipocyte TG hydrolysis can be stimulated by activators of protein kinase A, such as isoproterenol (6). We next evaluated whether the elevated TG hydrolysis in EKO adipocytes could be further stimulated by isoproterenol. As shown in Fig. 2C, transplanted WT adipocytes were clearly able to respond to this regulatory stimulus and increased their TG hydrolysis nearly twofold. The elevated TG hydrolysis in transplanted EKO adipocytes was not further increased by treatment with isoproterenol.
Adipocytes isolated from EKO mice display significant changes in gene expression, including reduced adiponectin, peroxisome proliferator-activated receptor (PPAR)-γ, and caveolin mRNA levels and elevated mRNA levels for genes involved in fatty acid oxidation and for perilipin (19). We next analyzed expression of these genes in adipocytes isolated from transplanted adipose tissue to determine whether the changes described above reflect the abnormal in vivo milieu of EKO mice or the absence of adipocyte apoE expression. Figure 3 shows the fold change for the indicated mRNA level in EKO compared with WT transplanted adipocytes. There was no difference in PPARγ mRNA expression. mRNA levels for adiponectin, perilipin, and proteins involved in fatty acid oxidation (proliferator-activated receptor coactivator, carnitine palmitoyltransferase I, and acyl-CoA oxidase) were higher in transplanted EKO than WT adipocytes. There was no difference in medium-chain acetyl-CoA dehydrogenase levels. Caveolin-1 mRNA levels were significantly lower in transplanted EKO adipocytes. The results of these transplantation experiments indicate that the lower adiponectin and PPARγ mRNA levels previously measured in adipocytes freshly isolated from EKO mice likely reflect the abnormal metabolic in vivo environment, inasmuch as these levels are not different (PPARγ) or are increased (adiponectin) in EKO adipose tissue harvested from a WT milieu. The higher level of gene expression for enzymes in the fatty acid oxidation pathway and the lower caveolin levels previously measured in freshly isolated adipocytes from EKO mice, however, are also observed in adipocytes isolated from EKO adipose tissue harvested from WT recipients and, therefore, reflect the absence of endogenous adipocyte apoE expression. There was no difference in mRNA abundance for lipoprotein lipase between transplanted WT and EKO adipocytes. mRNA levels for adipocyte TG lipase and hormone-sensitive lipase were significantly higher in transplanted EKO adipocytes, consistent with the higher rates of TG hydrolysis observed in Fig. 2B.
Table 1 shows TC, TG, and apoE levels in the two groups of recipient mice. There was no significant difference in TC or apoE level, but TG levels were ∼30% lower in the recipient mice that received EKO adipose tissue. This difference in TG level between WT recipients of EKO and WT adipose tissue was unexpected and raised two competing explanations for the results in Figs. 1–3: 1) endogenously produced apoE by adipocytes modulates adipocyte TG accumulation by its cell of origin, and 2) EKO adipose tissue secreted a factor(s) that modulated systemic lipid metabolism, resulting in altered lipoprotein composition and metabolism, as reflected by changes in the level of circulating TG and apoE, with a resultant decrease in adipocyte size. Our previously published in vitro observations (19) would favor the former explanation. To test these competing hypotheses, we 1) used an HFD to assess the response of transplanted adipocytes to increases in circulating TG and apoE and 2) evaluated adipocyte size in endogenous host WT adipose tissue harvested from recipients of transplanted WT or EKO adipose tissue. For the first approach, after transplantation of EKO or WT adipose tissue, recipient WT mice were fed chow or HFD for 10 wk before harvest of adipose tissue. Table 2 shows TC, TG, and apoE levels in recipients of WT or EKO adipose tissue on each diet. Within each transplantation group, TC and apoE levels increased significantly in the HFD-fed groups. TG levels also increased in both HFD-fed transplantation groups, but changes did not reach statistical significance. On the HFD, TC and apoE levels were not different between the two transplant groups, but TG level was 22% lower in recipients of the EKO adipose tissue. A representative section of transplanted EKO and WT adipose tissue harvested from HFD-fed WT recipients in Fig. 4A demonstrates the larger adipocytes in transplanted WT adipose tissue. A frequency plot of adipocyte volume from transplanted adipose tissue harvested from HFD-fed recipients is shown in Fig. 4B. Mean volume for transplanted EKO and WT adipocytes from recipients fed the HFD or chow diet is shown in Fig. 4C. EKO adipocytes from chow- and HFD-fed recipients were significantly smaller. Although a difference of 22% in TG level was observed between recipients of EKO and WT adipose tissue fed the HFD, evaluation of the response of each type of transplanted adipocyte to the HFD shows that WT adipocytes responded to the HFD (with its attendant 62% increase in TC and apoE and 30% increase in TG) with an increase in adipocyte volume of 106 × 103 μm3 (∼2-fold, P < 0.01). The response of EKO adipocyte size to the HFD was much more modest (19 × 103 μm3, not significant), despite equal or greater increases in TC, TG, and apoE produced by the HFD. In addition, adipocyte volume remained significantly lower (by 54%) in transplanted EKO adipocytes harvested from HFD-fed recipients than transplanted WT adipocytes harvested from chow-fed recipients (Fig. 4C; P < 0.05), even though TC, TG, and apoE levels were the same or higher in the former group (Table 2). Analysis of gene expression in transplanted EKO and WT adipocytes harvested from HFD-fed recipients is shown in Fig. 5. Similar to chow-fed recipients (Fig. 3), levels of adiponectin, proliferator-activated receptor coactivator-1, carnitine palmitoyltransferase I, acyl-CoA oxidase, and perilipin were significantly increased, caveolin remained significantly lower, and there was no difference in PPARγ between transplanted EKO and WT adipocytes in HFD-fed recipients. Difference in expression level for medium-chain acetyl-CoA dehydrogenase became significant in the HFD-fed recipients.
We next used another approach to evaluate a direct effect of endogenous adipocyte apoE on lipid accumulation by its cell of origin. WT host mice received adipose tissue from WT or EKO donors and were fed the HFD. At the time of euthanization, transplanted adipose tissue, as well as endogenous WT recipient visceral and subcutaneous adipose tissue, was harvested for evaluation of adipocyte size. As shown in Fig. 6, top, consistent with results in Figs. 1 and 4, EKO adipocytes harvested from transplanted EKO adipose tissue were significantly smaller than WT adipocytes harvested from transplanted WT adipose tissue. On the other hand, in adipocytes isolated from recipient host WT adipose tissue, there was no difference in size between recipients of the different types of transplanted adipose tissue. This finding establishes that the smaller size of transplanted EKO compared with WT adipocytes was not due to differences in circulating TG level or any systemic effect of transplanted EKO adipose tissue.
The largest portion of adipocyte TG content in vivo depends on uptake of lipid from circulating TGRL. The above-described results are consistent with the hypothesis that endogenously produced adipocyte apoE facilitates lipid acquisition from circulating TGRL by its cell of origin. In the next series of experiments, we evaluated lipid acquisition, as indexed by adipocyte TG synthesis, by isolated adipocytes during incubation with apoE-containing TGRL. Freshly isolated EKO or WT adipocytes were incubated alone or with apoE-containing VLDL for 6 h (Fig. 7A). In the absence of VLDL, TG synthesis was lower in EKO adipocytes, consistent with the results in Fig. 1. The addition of apoE-containing VLDL increased TG synthesis in WT adipocytes by almost fivefold, whereas the response of EKO adipocytes was much more modest, whether considered as fold change or absolute increase. Similar results were obtained when EKO or WT preadipocytes were differentiated in culture and incubated alone or in the presence of apoE-containing VLDL (Fig. 7B). These results indicate that the absence of endogenously produced apoE impairs adipocyte TG acquisition from apoE-containing TGRL.
In this series of studies, we tested the hypothesis that endogenously produced apoE plays an important in vivo role for modulating adipocyte TG metabolism. Such a role is suggested by observations that mature adipocytes isolated from EKO mice are smaller and demonstrate lower rates of TG synthesis, higher rates of TG hydrolysis, and altered gene expression compared with those isolated from WT mice (5, 10, 19). These previous observations, however, could not distinguish between the impact of the absence of adipocyte apoE expression and that of a profoundly disturbed in vivo environment resulting from systemic absence of apoE. The hypothesis that endogenous adipocyte apoE plays an important role in adipocyte lipid metabolism and gene expression is supported by in vitro studies using cultured adipocytes and adipose tissue (19). The present studies indicate that when EKO adipose tissue is placed in an in vivo WT environment, EKO adipocytes demonstrate significant differences in lipid metabolism and gene expression compared with WT adipocytes. This result could be produced if the developmental absence of apoE leads to permanent defects in adipocyte homeostasis. Our previous in vitro results showing that adenoviral expression of apoE in EKO adipocytes restores adipocyte TG metabolism to normal, however, favor the conclusion that ongoing adipocyte apoE expression is an important regulator of adipocyte differentiated function. The results of the present study importantly extend the physiological relevance of adipocyte apoE expression and establish a novel in vivo role for endogenously produced apoE in adipocyte size, TG metabolism, and gene expression.
The observation that recipients of EKO adipose tissue tended to have lower circulating TC, TG, and apoE levels than recipients of WT adipose tissue was unexpected and has not been observed in all transplantation experiments (Z. H. Huang and T. Mazzone, unpublished observations). The precise mechanism accounting for this observation remains to be determined in ongoing and future studies. However, to address our hypothesis that endogenously produced apoE modulates TG metabolism by its cell of origin in vivo, we needed to understand whether differences in circulating lipids contributed to the differences we observed in transplanted adipose tissue. We addressed this issue using two experimental approaches: 1) by using an HFD to increase circulating lipids and apoE and evaluating the effect of these changes on adipocyte size and 2) by evaluating adipocyte size in the endogenous adipose tissue of transplant recipients. The volume of transplanted WT adipocytes almost doubled in response to the HFD. On the basis of the observations of others (5, 10), this could be related to the 62% increase in TC and apoE levels and the 30% increase in TG levels. The change in volume of EKO adipocytes was much more modest in response to larger increases (92%, 46%, and 79% in TC, TG, and apoE levels, respectively) in the HFD-fed recipients. These results demonstrate that the endogenous production of apoE facilitates adipocyte lipid accumulation in response to the increase in circulating lipids and apoE produced by the HFD. In addition, when compared in the same experiment, transplanted EKO adipocytes harvested from HFD-fed recipients were significantly smaller (by 54%) than WT adipocytes harvested from chow-fed recipients, despite comparable lipid and apoE levels. Most importantly, adipocyte size in endogenous host adipose tissue is not different between recipient animals transplanted with EKO and those transplanted with WT adipose tissue. This latter observation rules out the possibility that a modestly lower TG level in recipients of transplanted EKO adipose tissue contributes to smaller EKO adipocyte size and supports the conclusion that the absence of endogenously produced adipocyte apoE impairs acquisition of lipid by its cell of origin. Because most adipocyte lipid in vivo is derived from circulating TGRL, our results support a mechanism whereby endogenously expressed apoE facilitates adipocyte acquisition of lipid from circulating TGRL. Consistent with this, both freshly isolated EKO adipocytes and those maintained in culture synthesize significantly less TG during a 6-h incubation with apoE-containing TGRL. In Fig. 2A, TG synthesis was measured using labeled oleate as a marker; in Fig. 7, TG synthesis was determined after adipocytes were labeled with glucose and its incorporation as glycerol into TG was measured. The identical results using these two complementary methods indicate that endogenous expression of apoE could increase adipocyte TG content by facilitating internalization of free fatty acid derived from extracellular lipolysis of TGRL or by enhancing TGRL particle uptake.
Our data also demonstrate differences in adipocyte gene expression as a result of the absence of apoE expression. It is possible that many of these changes result from downstream effects of apoE on adipocyte lipid flux. For example, apoE has an important role in sterol flux in multiple cell types (1, 23), and levels of caveolin mRNA may respond to cell sterol (13). Altered TGRL and fatty acid flux could impact the generation of intracellular ligands for PPAR family members that are importantly involved in regulating adipocyte gene expression (25, 27, 31). For example, activation of PPARδ in adipose tissue has been shown to increase the expression of genes involved in free fatty acid oxidation (31). The elevation of perilipin mRNA levels in transplanted EKO adipocytes is unexpected in view of their smaller size and lower lipid content. However, the impact of altered apoE expression on perilipin expression and/or function is complex, as indicated by our previous observations that the increase in perilipin mRNA levels in response to the adipogenesis stimulated by PPARγ agonists is attenuated in EKO adipocytes (19). Furthermore, because perilipin is a member of a family of proteins with overlapping function and is subject to significant posttranslational regulation, the relationship between perilipin expression and adipocyte TG metabolism remains unsettled in the literature (6).
High-level expression of apoE in adipocytes was first noted by Zechner and colleagues (34), and additional information regarding physiological regulators of adipocyte apoE has recently become available. Systemic administration of PPARγ agonists or in vitro incubation with these agonists increases adipocyte apoE expression, and TNFα and reactive oxygen species markedly suppress its expression (18, 32, 33). Obesity, either diet induced or leptin deficient, produces a marked reduction in adipocyte apoE, and important regulatory signals for reducing adipocyte apoE in obesity originate in the adipose tissue stromovascular fraction (7, 18). Weight loss associated with acute food deprivation or more prolonged fasting is associated with elevated adipocyte apoE (7). This pattern of physiological regulation, along with in vitro and in vivo observations relating to the function of adipocyte apoE in adipocyte lipid metabolism and gene expression, supports an important role for adipocyte apoE in modulating adipocyte response to metabolic and nutritional signals. The impact of endogenously produced apoE for regulating adipocyte TG homeostasis suggests that factors that regulate adipocyte apoE expression could influence the partitioning of circulating TGRL lipid between adipose and nonadipose tissues.
This work was supported by a gift from the Svendsen Family Foundation and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-71711 (to T. Mazzone).
The authors thank Stephanie Thompson for assistance with manuscript preparation.
- Copyright © 2009 the American Physiological Society